Trajectory Clustering Method Based on Multi-clustering Results Merging
Han Chen-shou · Microelectronics & Computer · 2011
In view of the unreliable of trajectory clustering results,a trajectory clustering method based on multi-clustering results merging(MRMTC) is proposed in this paper.For the representative trajectories of multi-clustering generated by clustering devices,a trajectory merging algorithm is proposed to merge them.The merging algorithm uses average scan line distance function as consensus function to compare similarities of representative trajectories,and then merges the similar representative trajectories.Finally,experiments results show that the proposed method MRMTC can produce more stable and effective clustering results.